• 1 Introduction

  1.1 Political Background

  The general attitude towards violent and sex offenders changed in Germany in the second half of the 1990s. The reasons were multifaceted, but of major importance was the media coverage of several sexual offences against young girls in the mid-1990s. There was the much talked about case of Marc Dutroux, who kidnapped, raped and killed teenage girls in Belgium. In Germany, there was the case of a seven-year-old girl who was abused and murdered by a patient during day release in the vicinity of a mental hospital in 1994. There were the murders of two girls in 1996 and 1997 in Bavaria and Lower Saxony who were sexually abused and killed by ex-convicts after their release from a prior prison sentence. The call for harsher sentences for ‘dangerous’ sex offenders grew louder and exerted considerable pressure on the German legislature, the judiciary, and the authorities of prisons and mental hospitals.
  Parliament reacted in 1998 by the enactment of a law that brought remarkable changes to the German system of criminal sanctions.¹x Act to combat sexual offences and other dangerous offences of 26 January 1998, Federal Law Gazette I page 160; the act is commented by H. Schöch, ‘Das Gesetz zur Bekämpfung von Sexualdelikten und anderen gefährlichen Straftaten vom 26.1.1998’, 51 Neue Juristische Wochenschrift 1257 (1998). The system as a whole still dated from a reform period in the late 1960s and had undergone the lapse of time without major interventions by public policy. The new law tightened the requirements for conditional release from prison and other custodial institutions by accentuating the necessity to consider public security interests. It became obligatory to obtain the opinion of a psychiatrist before a long-term sentence or a custodial preventive measure could be suspended. Directions to undergo medical treatment were allowed without the prior consent of the convicted person as long as he was not of an invasive nature or treatment for addiction, which meant that a convicted person could be compelled to go to psychotherapeutic treatment after his release. The scope of post-release supervision was widened; it was imposed by law on sex offenders having served a full sentence of more than one year in prison, whereas the law stipulated at the same time that other offenders must have served more than two years to trigger supervision after release. The courts were allowed to impose indeterminate supervision, if the offender did not consent to or comply with a medical treatment or an addiction treatment direction. Last but not least, the scope of preventive detention, in place since the 1930s, was widened by reducing the formal requirements for its imposition if an offender had been convicted of a felony (minimum prison sentence of one year), a sex offence or a qualified offence against the person.
  The general attitude was not softened by these changes, fear of crime and the willingness to inflict punishment on dangerous offenders, namely sex offenders, remained on a high level. Although international comparisons showed that public opinion in Germany differed considerably from that in other countries, especially England and Wales and the USA, in that the approval rate of the public for imprisonment was still remarkably lower in Germany than in other countries,²x J. van Dijk, J. van Kesteren & P. Smit, Criminal Victimisation in International Perspective. Key Findings from 2004 – 2005 ICVS and EU ICS (2007), at 147; K.-H. Reuband, ‘Steigende Repressionsneigung im Zeitalter der “Postmoderne”?’, 15 Neue Kriminalpolitik 100 (2003); K.-H. Reuband, ‘Steigende Punitivität oder stabile Sanktionsorientierungen der Bundesbürger? Das Strafverlangen auf der Deliktebene im Zeitvergleich’, 21 Soziale Probleme 97 (2010). the general attitude could not be ignored by German policy makers. Not only the conservatives who had authored the law of 1998, but also the social democrats who were in power 1998-2005 stood for a rigid course on dangerous and sex offenders. In 2001, Chancellor Schröder publicly articulated what many people deemed necessary: ‘Lock them away, forever’.³x E.C. Rautenberg, ‘Wegschließen für immer!?’, 54 Neue Juristische Wochenschrift 2608 (2001). As a consequence, the first decade of the new century saw a continued shifting in the system of criminal sanctions with a further tightening of the restrictions on dangerous and sex offenders. Again and again, the scope of preventive detention was extended. The limitation of ten years for the first-time detainees was abolished retroactively, new forms of deferred and subsequent detention were introduced, and unlimited preventive detention was allowed for young adults and even for juveniles.⁴x J. Kinzig, Die Legalbewährung gefährlicher Rückfalltäter. Zugleich ein Betrag zur Entwicklung des Rechts der Sicherungsverwahrung (2008), at 9. In addition, but without reference to specific incidents or cases, the provisions of supervision were enhanced and adapted to new security requirements in 2007.⁵x Act to reform supervision and to revise the provisions on subsequent preventive detention of 13 April 2007, Federal Law Gazette I page 513, with commentary by U. Schneider, ‘Die Reform der Führungsaufsicht’, 27 Neue Zeitschrift für Strafrecht 441 (2007). Core elements of the Act of 2007 were the extension and fortification of the directions that may be imposed on a convicted offender after release as well as the further extension of the scope of indeterminate supervision. New directions were the order to present himself at certain times to a doctor, a psychotherapist or a forensic ambulance service, and the order to undergo psychiatric, psycho- or socio-therapy. Indeterminate supervision was additionally allowed for offenders, especially sex offenders, who violated directions, if this gave reason to believe that there was a danger to the general public by the commission of further serious offences.
  The year 2009 brought a turning point. On 17 December 2009, the European Court of Human Rights ruled that the 1998 retrospective extension of preventive detention to an unlimited period of time had been incompatible with the right not to have a heavier penalty imposed than the one applicable at the time of the offence (Art. 7 ECHR).⁶x ECtHR, M. v. Germany, Applic. No. 19359/04. On 13 January 2011, the Court confirmed that judgement and held that the subsequent order of preventive detention did not have the necessary causal connection between the conviction and the deprivation of liberty at issue, and therefore violated the detainee’s right to liberty and freedom (Art. 5 ECHR).⁷x ECtHR, Haidn v. Germany, Applic. No. 6587/04. On 4 May 2011, the German Federal Constitutional Court abandoned earlier judgements and found that there was no enough distance between the execution of a prison sentence and the execution of preventive detention, i.e. sentences and preventive detention were executed in similar settings and under similar conditions (e.g. in similar-sized cells). Therefore, all legal provisions of the imposition and duration of preventive detention were declared incompatible with the fundamental rights guaranteed by the German constitution, the so-called Basic Law.⁸x Federal Constitutional Court of Germany, 4 May 2011, 2 BvR 2365/09. Consequently, two new laws were enacted in 2010 and 2012, by which the requirements for the imposition and execution of preventive detention were redesigned. The first was a reaction to the 2009 decision of the European Court of Human rights; it practically abolished the subsequent order of preventive detention.⁹x Act to reorder the provisions on preventive detention of 22 December 2010, Federal Law Gazette I page 2300. The second was a reaction to the 2011 decision of the Federal Constitutional Court; it determined that the execution of preventive detention and the antecedent imprisonment have to aim at the minimisation of the offender’s dangerousness and offer enough measures for an open, liberty-oriented detention regime.¹⁰x Act to implement the imperative of distance in the provisions on preventive detention of 5 December 2012, Federal Law Gazette I page 2425. Since then, the general debate on the proper handling of dangerous offenders has become less agitated and fearful in Germany. Public opinion seems to have eased off for a while, at last.

  1.2 Supervision in the Twin Track System

  Before going into the details of the indeterminate control of sex offenders in the community, it should be noted that the German system of criminal sanctions is a twin track system like in many other countries on the European continent. ‘Twin track’ means that the divergent functions of criminal sanctions are separated in the sentencing process and assigned to either the retributive track of penalties (imprisonment and fines), or the preventive track of custodial or non-custodial measures (‘measures of rehabilitation and security’). Entry requirement for the first track is the offender’s legal responsibility, and that for the second is the likelihood of further offenses, which may be considered a synonym for the offender’s ‘dangerousness’. An offender who is convicted for drink-driving, for instance, is fined (punishment) and additionally disqualified from driving if necessary (a preventive measure) in Germany. A mentally ill or an addicted offender may receive a prison sentence and be additionally sent to a mental hospital or custodial addiction treatment, the time spent in custody for the latter credited against the sentence. A violent or a sex offender, to mention a third example, who is considered a danger to the general public, because the court finds a high risk of further serious offences resulting in serious emotional trauma or physical injury to the victim, has to serve the sentence of imprisonment first and is placed in preventive detention afterwards until the preventive measure is suspended or terminated. The advantage of the twin track system is seen in the effect that the penalties do not necessarily have to serve preventive needs. Because it is usually only a small group of offenders who show specific preventive needs, the consequence of the twin track system is that the general level of punishment can be kept at a rather moderate level (in Germany, the prison population rate per 100,000 population was 84.1 in 2013).¹¹x Council of Europe, Annual Penal Statistics. Space I – Prison Populations. Survey 2013 (2014), at 42. It is generally not assumed that the rate of recidivism, which appears to be quite high in Germany (46.2% of the persons released from imprisonment in 2007 were reconvicted after three years¹²x J.-M. Jehle, H.-J. Albrecht, S. Hohmann-Fricke & C. Tetal, Legalbewährung nach strafrechtlichen Sanktionen. Eine bundesweite Rückfalluntersuchung 2007 bis 2010 und 2004 bis 2010 (2013), at 36, available at: <www.bmjv.de/DE/Ministerium/Fachthemen/AbtII/IIA7/Rueckfallstatistik_doc.html?nn=1468684> (last visited 12 August 2015). An overview of recidivism studies in various European countries is given on the Internet: <http://wp.unil.ch/space/publications/recidivism-studies/27/> (last visited 19 April 2016). ), can effectively be reduced just by longer or harsher sentences.
  In the German twin track system of criminal sanctions, there are two differing forms of non-custodial control of offenders: probation and supervision. Probation is a part of the retributive, penalty track and dependent on the suspension of a prison sentence. In Germany, sentences not exceeding two years may be suspended if there are reasons to believe that the convicted offender will not commit further offences. The normative assumption is that, in these cases, the preventive needs are better met by avoiding the offender’s imprisonment. In the same way, the court may grant conditional early release from a prison sentence if the release is appropriate considering public security interests, i.e. if, again, it can be expected that the convicted person will not commit further offences. The basic requirement for probation is thus a low risk of reoffending, the courts having to be convinced that the offender will observe the directions, follow the guidance of a probation officer and not disappoint the expectation that no further offences are committed. The operational period may not exceed five years, an indeterminate control is not possible.
  Supervision, on the other hand, is a non-custodial preventive measure of rehabilitation and security. Either it may be imposed by the courts, which is not relevant in practice (less than hundred cases per year¹³x Statistisches Bundesamt, Rechtspflege. Strafverfolgung (2013), table 5.5. ), or it emerges as the automatic consequence of the termination of a prison sentence or the suspension or termination of a custodial preventive measure (mental hospital, addiction treatment, preventive detention). The risks of further offences are varied in these cases. The risk is low if a custodial measure is suspended (because the suspension is bound by law to a low risk of re-offending, according to sec. 67d(2) German Criminal Code the offender must be expected not to commit any more unlawful acts if released). It is high if the offender had to serve a full prison sentence or if a measure is terminated because the maximum period has expired (addiction treatment, two years) or because the continued enforcement of the measure would be disproportionate; if it were low, the offender would have been released earlier. During supervision, the released person is under the guidance and control of a supervising authority, a probation officer, forensic ambulance services, the courts and in some cases the police.¹⁴x B.-D. Meier, Strafrechtliche Sanktionen, 4th edn (2015), at 296-301. The offender’s conduct is structured and influenced by directions imposed by the courts. The period of supervision may last between two and five years, or may be indeterminate. In all of these cases, supervision is terminated if the court expects that the convicted person will, even in the absence of further supervision, not commit any further offences.
  It is because of the last mentioned indeterminate, potentially life-long control that supervision is considered the most intrusive community measure of German law. In spite of its substantial theoretical significance, however, there are no official statistics on the frequency of supervision in Germany. The only data available stem from an unofficial gathering organised by Deutsche Bewährungshilfe e.V., a professional association for probation officers. This data collection indicates that there has been a considerable increase in supervision cases from 24,800 in 2008 to 36,700 in 2014 (+47.9%).¹⁵x See <www.dbh-online.de/fa/> (last visited 31 July 2015). The reasons for the rise are unknown, but it can be assumed that there is a connection with the development of mental hospital and custodial addiction treatment orders, which has also seen a considerable increase in the past two decades.¹⁶x M.P. Rohrbach, ‘Die Verschärfung der Führungsaufsicht’, 97 Monatsschrift für Kriminologie und Strafrechtsreform 248 (2014); see also C. Morgenstern, A. Hecht, ‘Rechtstatsachen zur Führungsaufsicht im kriminalpolitischen Kontext’, 58 Bewährungshilfe 186 (2011).

• 2 Indeterminate Supervision in the Light of Sentencing Theory

  In Germany, the origins of the twin track system date back to the scholarly discussions of the late nineteenth/early twentieth century on the justification of punishment. Two conceptions competed, the conception of punishment as retribution for a crime committed in the past (classical school, in Germany tied to the name of Karl Binding, 1841-1920) and the conception of punishment as a means to prevent future crimes (modern school, propagated by Franz von Liszt, 1851-1919).¹⁷x T. Vormbaum, Einführung in die moderne Strafrechtsgeschichte, 2nd edn (2011), at 137. The result of the dispute was a compromise, which took suggestions of the Swiss professor Carl Stooss (1849-1934) and split the differing functions of punishment into the two systematically different tracks of penalties and measures. The first draft law based on this compromise was published as early as 1909,¹⁸x Reichsjustizamt, Vorentwurf zu einem Deutschen Strafgesetzbuch. Begründung. Allgemeiner Teil (1909), at 147. but it was only a quarter of a century later that the new twin track system was formally enacted in 1933.¹⁹x Act against dangerous habitual offenders and on measures of security and rehabilitation of 24 November 1933, Imperial Law Gazette I page 995. Since then, the existence of the twin track system has become an integral and undisputed part of criminal law in Germany.
  In the German twin track system, retributive penalties are justified by looking back to the committed offence. The offender is punished because he has committed a crime, no further justification needed. At bottom, punishment is retaliation. It is a necessary means to the restoration of justice because, in retributive thinking, the wrong committed by the offender can only be evened out by the infliction of a suffering on him, a consideration unfolded by German idealism, mainly by Immanuel Kant (1724-1804). In today’s German criminal law theory, however, it is not the wrong of the offence that is counterbalanced by the sentence but rather the offender’s guilt, i.e. his personal responsibility for the violation of law. An offender may only be punished if when committing the crime he was aware that he was acting unlawfully, he was capable to appreciate the unlawfulness of his actions and act accordingly, and it could reasonably be expected from him to abide by the law because he was not in a situation of duress. Guilt is, in other words, the blameworthiness of the offender for not using his capacity to abide by the law.²⁰x Federal Court of Justice, 18 March 1952, GSSt 2/51.
  In sentencing theory, these considerations legitimising the right to punish are transferred to the determination of the severity of the sentence for a given offence. All aspects that can be related to the offence are evaluated in the light of guilt, which includes not only the offence itself but also the pre- and post-offence behaviour (earlier sentences, confession, redress, etc.).²¹x Meier, above n. 14, at 192; see also F. Streng, Strafrechtliche Sanktionen, 3rd edn (2012), at 255. It is in this sense the notion of guilt has to be understood if German law stipulates in sec. 46(1) Criminal Code: ‘The guilt of the offender is the basis for sentencing’. The popular saying that punishment should fit the crime thus does not apply in German law. Punishment is rather construed as a function of guilt. The concept of guilt does not imply that preventive aspects from deterrence to rehabilitation are completely irrelevant in the retributive track of punishment. They are not of major importance, though, and may only be considered in the range defined by the severity of offender’s individual guilt. A convicted person may therefore not be sent to prison for a longer period of time than required by the degree of his guilt, even if security reasons require the placement in a mental hospital or preventive detention afterwards.²²x Federal Court of Justice, 4 August 1965, 2 StR 282/65. As supervision, on the other hand, is not considered a form of punishment but a preventive measure, it may be inflicted without regard to the offender’s severity of guilt.
  The theoretical justification of the preventive track of custodial and non-custodial measures is less clear. Most authors identify a utilitarian principle of predominant public interest that legitimises any form of intervention into the private sphere as long as the safeguards of proportionality are observed.²³x W. Frisch, ‘Die Maßregeln der Besserung und Sicherung im strafrechtlichen Rechtsfolgensystem’, 102 Zeitschrift für die gesamte Strafrechtswissenschaft 369, at 382 (1990); G. Kaiser, Befinden sich die kriminalrechtlichen Maßregeln in der Krise? (1990), at 12; see also Federal Constitutional Court of Germany, 4 May 2011, 2 BvR 2365/09. In the case of preventive criminal sanctions, the principle of predominant public interest emerges as a means to protect the public against the danger of further offences. The offence committed in the past provides the motivation, but not the justification for the offender’s conviction and the imposition and execution of sanctions; the justification is derived solely from the risk of future offences and the imminent danger for public security. The principle of predominant public interest is, however, strictly bound to the proportionality of the sanction. As sec. 62 of the German Criminal Code states: ‘A measure of rehabilitation and incapacitation must not be ordered if its use is disproportionate to the seriousness of the offence committed by or expected to be committed by the convicted person and to the degree of danger he poses to society’. As sec. 62 applies to all forms of preventive measures in German law, supervision is likewise bound to the principle of proportionality. Supervision may only be imposed and executed if and as long as it is necessary to achieve its aim of public protection.
  In the focus of all preventive measures is the concept of danger, i.e. the prospect that further offences will be committed. Under the principle of proportionality, ‘danger’ must be understood as a dynamic concept, which needs to be adapted to the depth of intrusion into the private sphere going along with the different forms of preventive measures. Custodial measures may, for instance, be justified only with a greater degree of ‘danger’ than non-custodial measures, an indeterminate mental hospital order requires a greater degree of danger than a custodial addiction treatment order, which is limited to two years in German law.²⁴x A. Dessecker, Gefährlichkeit und Verhältnismäßigkeit. Eine Untersuchung zum Maßregelrecht (2004), at 132. Similar considerations apply to different embodiments of supervision. The more intensive forms, particularly indeterminate supervision or its combination with specific directions such as electronic monitoring, require a greater degree of danger than minor forms that may be less intrusive. The ‘degree of danger’ addressed in sec. 62 German Criminal Code can, at least in theory, be related to the likelihood as well as to the seriousness of the future offences; the more serious the expected offences are, the lower the likelihood by which they are expected may be, and vice versa. Nearness of the expected offences and the validity of the instruments used for risk assessment may be seen as additional aspects that may be taken into account. Proper judgements on the ‘danger’ a convicted person poses to society are thus quite a difficult task and sometimes hard to achieve in practice.²⁵x Ibid., at 187. Whenever a court feels that it lacks the necessary empirical knowledge, expert opinion must be heard.
  Supervision is construed as a preventive measure that is imposed and executed under criminal law in Germany. It is justified by the predominant interest in public security and is bound to the principle of proportionality, the hardship going along with its imposition or execution being acceptable, provided that it is proportionate to the degree of danger attributed to the offender. Indeterminate supervision is no exception. According to sec. 68c(2) German Criminal Code, indeterminate supervision is permitted, if the convicted person does not consent to or comply with a medical treatment or an addiction treatment direction and ‘a danger to the general public through the commission of further serious offences is to be expected’ (introduced in 1998). According to sec. 68c(3), indeterminate supervision is permitted, if in the case of a suspended mental hospital order there is reason to believe that the convicted person is otherwise about to lapse into a state of mental incapacity or if a violent or a sex offender violates directions imposed on him by the court or ‘other specific circumstances give reason to believe that there may be a danger to the general public by the commission of further serious offences’ (introduced in 2007). As indeterminate supervision results in a substantial, potentially life-long restriction of the offender’s liberty, the legal provisions have to be applied carefully and restrictively by the courts.²⁶x Marburg Regional Court, 13 January 2014, 11 StVK 9/14; K.-H. Groß, Münchener Kommentar zum Strafgesetzbuch, 2nd edn (2012), sec. 68c n. 7; J. Kinzig, Schönke/Schröder. Strafgesetzbuch. Kommentar, 29th edn (2014), sec. 68c n. 2.
  There is, of course, more academic debate on sentencing theory, the principles of sentencing, the theoretical justification of the preventive measures, and the concepts of danger and proportionality than has been outlined here. The concept of personal guilt has been challenged by authors who recommend the replacement by a sentencing system that is solely built on the severity of the offence and the concept of harm, thus unburdening the present system of the necessity to consider highly individualised aspects of personal guilt and preventive needs.²⁷x T. Hörnle, Tatproportionale Strafzumessung (1999); T. Hörnle, ‘Das antiquierte Schuldverständnis der traditionellen Strafzumessungsrechtsprechung und –lehre’, 54 Juristenzeitung 1080 (1999); but see also Streng, above n. 21, at 314-17. Moreover, it has been pointed out that the concept of danger, which underlies all preventive measures, requires reliable forensic prediction methods independent of judgemental discretion, which are difficult to achieve in practice; the suggestion has been that the empirical foundation of the preventive measures should be replaced by a normative system of a fair distribution of risks.²⁸x Streng, above n. 21, at 166-67; W. Frisch, ‘Defizite empirischen Wissens und ihre Bewältigung im Strafrecht’, in R. Bloy, M. Böse, T. Hillenkamp, C. Momsen & P. Rackow (eds.), Gerechte Strafe und legitimes Strafrecht. Festschrift für Manfred Maiwald (2010), 239 at 252; W. Frisch, Prognoseentscheidungen im Strafrecht. Zur normativen Relevanz empirischen Wissens und zur Entscheidung bei Nichtwissen (1983), at 53-55, 161-70. All of these approaches, however, have not been taken up by the legislature or the judiciary; their influence has been limited to academic circles. Nevertheless, in the past few years, it could be observed that there has been a new, increased awareness of the methodological problems of prediction methods throughout the courts and the forensic institutions.

• 3 Indeterminate Supervision and the Principle of Proportionality

  3.1 Law in Books

  In German law, the principle of proportionality consists of four elements: there must be an aim for the measure legitimate under Basic Law, the measure must be suitable to achieve the aim, the measure must be necessary to achieve the aim, i.e. the aim cannot be reached by less intrusive means, and considering the competing interests the measure must be adequate.²⁹x J. Kaspar, Verhältnismäßigkeit und Grundrechtsschutz im Präventionsstrafrecht (2014), at 101; Dessecker, above n. 24, at 343. In the context of supervision, the first of these elements is not questionable. It is generally acknowledged that the protection of the public is an objective that may legitimately be pursued by the state. Averting dangers to protect the public is in Germany considered one of the main functions of the state, which may even be derived from the fundamental rights guaranteed by the constitution. According to the German Federal Constitutional Court, the basic rights do not only constitute a defence against the state, but they also establish claims, i.e. in the case of criminal law, the claim to receive protection against unlawful attacks.³⁰x Federal Constitutional Court of Germany, 25 February 1975, 1 BvF 1/74; 29 May 1993, 2 BvF 2/90; K. Stern, ‘Die Schutzpflichtenfunktion der Grundrechte: Eine juristische Entdeckung’, 63 Die öffentliche Verwaltung 241 (2010); V. Epping, Grundrechte, 6th edn (2015), at 59.
  The second element, suitability, depends on how the measure of supervision is organised. It is obvious that the public protection provided by a non-custodial measure differs from the protection granted by custodial measures such as preventive detention or imprisonment. Nevertheless, supervision is not a totally inept (and therefore disproportionate) means to prevent the convicted person from re-offending. German law uses a twofold approach for the execution of supervision, which is a combination of control and support. Most powerful is the supervising authority, a branch of the general administration of justice that has the task to control the conduct of the convicted person and the fulfilment of the directions imposed upon him by the court.³¹x A. Dessecker, ‘Von der Polizeiaufsicht zur Führungsaufsicht und zurück?’, in E. Hilgendorf and R. Rengier (eds.), Festschrift für Wolfgang Heinz (2012) 631, at 634. If the convicted person does not comply with the directions, the supervising authority informs the court and has the right to file a request for prosecution because non-compliance with court orders constitutes a new offence in Germany punishable with up to three years of imprisonment (sec. 145a German Criminal Code). Social support, on the other hand, comes from a probation officer who is assigned by the court. The probation officer has the task to offer assistance and care according to social work standards. Their work is supplemented by forensic ambulance services, which were introduced in 2007. If the court imposed a therapy direction, i.e. a direction to undergo psychiatric, psycho- or socio-therapy, the forensic ambulance services offer these options to the offender. The three institutions (supervising authority, probation officer and forensic ambulance service) are by law obliged to work together and disclose secrets to each other (sec. 68a German Criminal Code). In the case of sex offenders released from prison or custodial preventive measures, the cooperation of the three institutions is additionally reinforced by the police. Since 2006, all German states have developed special programmes for the supervision of this particular group of offenders. The common feature of these programmes is that all sex offenders undergo individual risk assessments in round-table conferences chaired by the police. In these conferences, the risk is categorised, and the measures considered necessary and appropriate to prevent the convicted sex-offenders from re-offending are determined individually.³²x D. Ruderich, Führungsaufsicht (2014), at 245; S. Thomaßen, ‘Konzeption zum Umgang mit rückfallgefährdeten Sexualstraftätern (KURS NRW)’, 6 Forensische Psychiatrie, Psychologie, Kriminologie 25 (2012); M.P. Rohrbach, Die Entwicklung der Führungsaufsicht unter besonderer Berücksichtigung der Praxis in Mecklenburg-Vorpommern (2014), at 152; Rohrbach, above n. 16, at 257. Although the effectiveness of these programmes or of supervision as a whole has not yet been evaluated empirically in Germany, it can safely be assumed that the measures help to reduce the risk of further offending and thus make a contribution to public protection. The measure of supervision is, in other words, not unsuitable to reach its aim.
  The third element of proportionality, necessity, is dependent on the existence of less intrusive means to achieve the aim of public protection. In the German legal system, these less intrusive means can only come from the police laws, i.e. from the laws addressing safety and security needs and requirements. In this context, it is necessary to understand that in the German legal system, the position of the police is based on two foundations: criminal law and security law as part of administrative law. If police activities result from a criminal offence committed in the past, the legal basis is criminal law; in all other cases it is security law.³³x F. Schoch, ‘Doppelfunktionale Maßnahmen der Polizei’, 35 Juristische Ausbildung 1115 (2013). Most preventive measures imposable under criminal law can be imposed under the security laws as well if a danger for public security is expected. Football hooligans, for instance, which may not have committed a criminal offence in the past, may nevertheless be registered, categorised, addressed and supervised according to the security laws if the police expects a danger for public security, especially if they expect the commission of an offence.³⁴x T. Siegel, ‘Hooligans im Verwaltungsrecht. Stadionverbote und andere polizeirechtliche Maßnahmen zur Eindämmung von Gewalt in Fußballstadien’, 66 Neue Juristische Wochenschrift 1035 (2013). Under specific circumstances, even arrest and detention of a hooligan may be justified and are compatible with European law.³⁵x ECHR, Ostendorf v. Germany, Applic. No. 15598/08. Police activities based on the security laws can thus not necessarily be considered less intrusive than activities based on criminal law, although they are bound to a different, more restrictive degree of ‘danger’: the police laws stipulate that the danger must be ‘concrete’, i.e. it must be sufficiently probable that public security will be affected by some damaging event. The preventive measure of supervision under criminal law, on the other hand, is only bound to the ‘danger that the (convicted) person will commit further offences’ (sec. 68(1) German Criminal Code), which appears to be a more general concept. The differences, however, are marginal. The criminal courts are by law obliged to terminate supervision, ‘if it can be expected that even in the absence of further supervision the person will not commit any further offences’ (sec. 68e(2) German Criminal Code). It is therefore hardly possible to claim that supervision based on criminal law is disproportionate because public security measures under police law are less intrusive.
  The last element of proportionality, adequacy, compares the different interests at hand. The hardship imposed upon the individual must be appropriate to the gains reached for the public. It is in this context that the intensity in type and duration of the particular form of supervision (e.g. its indeterminate length) must be weighted and balanced against the likelihood and the seriousness of further offences that the convicted person is expected to commit in the absence of supervision, and the imposition of supervision is expected to avert. The seriousness of the offence(s) committed in the past, which is also mentioned in sec. 62 German Criminal Code as an element of the evaluation, is in this context only of minor importance because supervision is typically only imposed if the offence(s) committed in the past may be considered as ‘serious’; whether imposed by court or executed as an automatic consequence of the judgement, supervision becomes operational only in conjunction with custodial sanctions, not with fines or other non-custodial consequences of the committed offence. Provided that the degree of danger that is necessary for the imposition of supervision (notably the seriousness and the likelihood of the future offences) can be established in the particular case, the proportionality test thus requires that the depth of intrusion into the offender’s private sphere is elaborated and weighted. This can hardly be done in a general way because in practice supervision may be executed in a variety of forms emphasising either the aspects of social support and reintegration or the aspects of guidance and control. For a legal discussion, however, it is sufficient to know that supervision is established by law as a flexible instrument that can and must be adapted to the convicted person’s individual developments after release.³⁶x H. Schneider, Strafgesetzbuch. Leipziger Kommentar, 12th edn (2008), sec. 68d n. 6. All decisions on the particular execution of supervision can subsequently be reviewed, modified or vacated by the courts (sec. 68d(1) German Criminal Code), the maximum duration may be reduced and the measure may be terminated (sec. 68c(1) and 68e(2) German Criminal Code). All subsequent decisions are subject to legal review (sec. 463(2) German Code of Criminal Procedure). The flexibility and adaptability of the law thus preclude (at least in theory) a situation in which supervision may be considered a hardship for the offender not in balance with the gains reached for public security. From a theoretical point of view, supervision, and even indeterminate supervision, is consistent with the principle of proportionality.³⁷x For determinate supervision, see Federal Constitutional Court of Germany, 15 August 1980, 2 BvR 495/80.

  3.2 Law in Action

  The obvious question is how the theoretical concept of proportionality is applied in practice, i.e. whether the imposition and execution of supervision are applied in a way that is suitable, necessary and adequate to achieve the aim of public protection. Empirical data are available from a study conducted by a research team headed by German criminologists Jörg Kinzig and Alexander Baur from Tübingen University. The study is based on the examination of 606 records of supervision cases and the questioning of over 900 professionals in the field of supervision; it can be considered representative for the situation in Germany.³⁸x A. Baur and J. Kinzig (eds.), Die reformierte Führungsaufsicht (2015).
  Kinzig and Baur differentiate between three types of supervision: supervision after suspension of a custodial preventive measure (type I, low risk), supervision after termination of a custodial preventive measure (type II, which is a combination of different subgroups, high risk) and supervision after having served a full prison sentence (type III, high risk). In all of these cases, supervision is imposed by law and the courts’ decisions are restricted to the execution of the measure. The judicial decisions on the execution include assignment of the supervising authority and appointment of a probation officer, imposition of directions, determination of the period of supervision and subsequent decisions. With respect to the principle of proportionality, the determination of the duration of supervision is most interesting. Although the law provides that the period may be fixed between two and five years, five years being the preset duration, if the courts make no decision (sec. 68c German Criminal Code),³⁹x Dresden Higher Regional Court, 12 December 2008, 2 Ws 380/08; H. Ostendorf, Nomos Kommentar Strafgesetzbuch, 4th edn (2013), sec. 68c n. 5; but see also Berlin Higher Regional Court, 20 June 2011, 2 Ws 159/11. in practice the courts are reluctant to exert the discretion granted by law and in most cases stick to the maximum duration of five years instead. In the Kinzig and Baur study, the period was held at five years in 64.6% of the type I cases, 65.5% of the type II cases and 81.2% of the type III cases, whereas the minimum of two years was fixed only in 0.8%, 4.3% and 2.0% of the three groups, respectively.⁴⁰x Baur and Kinzig, above n. 38, at 533. This distribution is noticeable because it differs from the distribution that might be expected if probability calculation was applied; especially in the type I group of low-risk offenders, one might expect a remarkably lower percentage of maximum periods. Even if Kinzig and Baur also found that life-long supervision is in practice very rare,⁴¹x Ibid., at 538 and 543. the distribution thus signals that the majority of the courts act risk oriented and cautious in Germany, showing less interest in the offenders’ liberty rights than in the idea of public security. Apart from that, the theoretical option to revise the one-sided distribution by subsequent decisions (sec. 68d German Criminal Code) is in practice made use of only in a small number of cases.
  In light of these findings, the question arises, what is known about the ‘real’ rate of further offences committed by convicted persons under supervision? The effort to give an empirically validated answer is confronted with a series of methodological problems. Nevertheless, there is an analysis of the national register of criminal convictions conducted under the guidance of German criminologists Jörg-Martin Jehle and Hans-Jörg Albrecht. The study analysed inter alia the reconviction rates of all persons who had either been convicted and received a non-custodial sanction by German courts in 2007, or been released from imprisonment or a custodial preventive measure in 2007; in both cases the period under observation was three years. Directly comparable to the Kinzig/Baur study are only the results concerning convicted persons with a full prison sentence (type III supervision in the Kinzig and Baur study). Jehle and Albrecht found that, in this group, 40.4% of the persons released from prison re-offended within the observation period of three years. Other than in the Kinzig/Baur study in which the authors differentiated between suspension and termination of a custodial measure (type I and II), Jehle and Albrecht differentiated between offenders who had been sanctioned with an isolated custodial measure and offenders who had received the measure as an addendum to a retributive punishment (i.e. a combination of retributive and preventive punishment). In the first case, the reconviction rates were considerably lower (5.0% after release from mental hospital, 24.2% after release from custodial addiction treatment) than those in the second case (23.2% after mental hospital, 41.9% after addiction treatment, 39.4% after preventive detention).⁴²x Jehle et al., above n. 12, at 83. The reconviction rates in the Jehle and Albrecht study, however, include all kinds of convictions (offences similar to the earlier offences or not) and all kinds of sanctions (custodial and non-custodial); they may therefore not necessarily be interpreted as indicators of a general ‘dangerousness’ of the convicted persons under supervision. Reconvictions because of serious offences similar to the first offence are not as frequent as a misreading of the Jehle and Albrecht figures might suggest.
  Empirical data, in the contrary, suggest that the likelihood of further offences is overestimated in judicial practice. In Germany, several studies have been conducted on the development of convicted persons’ criminal behaviour after release. The examined persons either could not be sent to preventive detention or had to be released from preventive detention for formal reasons, although the judicial authorities were convinced that the persons in question were ‘dangerous’, i.e. although they exhibited a high risk of committing further serious violent or sex offences. A study by Jürgen L. Müller et al. found that of all offenders officially considered as ‘dangerous’ (n = 25), 60% re-offended in an observation period of two years. Only in 28% of the cases, however, the authors classified the re-offence as ‘serious’, because the offender received an unsuspended prison sentence of more than eighteen months, while 32% of the cases were considered ‘minor’.⁴³x J.L. Müller, ‘Legalbewährung nach Gutachten zur nachträglichen Sicherungsverwahrung’, 94 Monatsschrift für Kriminologie und Strafrechtsreform 256 (2011). Similar results were shown in a study by Michael Alex from Bochum University. In his study, Alex examined a sample of 121 high-risk offenders who had to be released from prison, after they had served their full sentence, although the public prosecutor still considered them dangerous. In this group, Alex found a rate of 52% of ex-inmates re-convicted after an observation period of at least forty-eight months. In more than half of the cases (29.8%), the re-offence was considered ‘serious’, because the offenders were sanctioned with unsuspended prison sentence, while in the rest of the cases the offenders received milder forms of punishment.⁴⁴x M. Alex, ‘Kriminalprognose und Legalbewährung – Wie zuverlässig lässt sich Rückfallgefahr vorhersagen?’, in K. Höffler (ed.), Brauchen wir eine Reform der freiheitsentziehenden Sanktionen? (2015) 21, at 26.
  These and similar results in other studies⁴⁵x Kinzig (2008), above n. 4, at 213 and 217. can be explained by the systematic difficulties hindering the exact prediction of events with a low base rate: the lower the base rate, the larger is the rate of false positives, if all other circumstances are kept equal.⁴⁶x J. Kühl and K.F. Schumann, ‘Prognosen im Strafrecht – Probleme der Methodologie und Legitimation’, 7 Recht & Psychiatrie 126 (1989). As serious offences have a low base rate (the more serious the lower it is), the prediction of future offences is burdened with the methodological problem that false positives will inevitably be predicted more often than false negatives. Prediction is, in other words, burdened with a systematic bias that seems to be unknown to the majority of courts but that becomes apparent in the results of empirical research.⁴⁷x Meier, above n. 14, at 372. Moreover, it must be noted that the factual basis of prediction is often quite thin. The future development of many of the risk and protective factors considered relevant by most prediction methods (e.g. relations to significant others, employment and place of residence) is unknown to courts and expert witnesses at the time of prediction, so their influence on the convicted person’s future behaviour can hardly be assessed properly.⁴⁸x Alex, above n. 44, at 33. The consequence is a bitter finding about the capacity of the principle of proportionality to guide and control judicial decisions in the field of preventive sanctions: if the prediction of the ‘degree of danger’ (cf. sec. 62 German Criminal Code), which is necessary for the imposition and execution of these sanctions, is in practice flawed by the inevitability of systematic errors, the protective function of the proportionality principle is in theory remarkably more promising than in practice. The necessity and the adequacy of (indeterminate) supervision can in practice easily be misjudged.

• 4 Alternative (Less Intrusive) Means

  It has already been pointed out that there are no means in German law that are as suitable as supervision to achieve the aim of public protection and that are at the same time less intrusive. Police law interventions are not necessarily less intrusive than post-release supervision. Civil law offers no alternatives as well. In 2010, a new law was enacted in Germany, which provides for the placement of persons who had to be released from preventive detention, because the retroactive application of indefinite post-sentence preventive detention was interpreted by the European Court of Human Rights as amounting to the retroactive imposition of a heavier penalty incompatible with European human rights law.⁴⁹x Act on the therapeutic placement of mentally disturbed violent offenders of 22 December 2010, Federal Law Gazette I page 2300. These persons may be placed in closed institutions if they are considered to be of unsound mind (cf. Art. 5(1)(e) ECHR) and if there is a high probability that they will affect the life, personal integrity, freedom or sexual self-determination of others so that their placement is considered necessary for public protection (sec. 1 Therapy and Placement of Mentally Disturbed Violent Offenders Act). The act was held compatible with German constitutional law, provided it is interpreted in accordance with the constitution meaning that there must be a very high risk of further crimes with a very severe impact on the victims.⁵⁰x Federal Constitutional Court of Germany, 11 July 2013, 2 BvR 2302/11. Placement in a closed institution thus is not a measure less intrusive than supervision and may therefore not be seen as an alternative. Neither may the appointment of a custodian to the convicted person be considered an alternative. According to German civil law, a custodian may be appointed if a person by reason of a mental illness or a physical, mental or psychological handicap cannot take care of his affairs (sec. 1896 German Civil Code). The aim of the appointment, however, is not public protection but the safeguarding of the best interests of the person under custodianship (sec. 1901(2) German Civil Code). The target groups thus are different; custodianship and supervision have objectives that must be kept apart clearly.
  Against this background, it may come as a surprise that there are some authors in German academic literature who are of the opinion that supervision is not compatible with the principle of proportionality.⁵¹x Ostendorf, above n. 39, pre secs. 68-68g n. 15; see also H. Pollähne, ‘Effektivere Sicherheit der Bevölkerung und schärfere Kontrolle der Lebensführung’, 90 Kritische Vierteljahresschrift für Gesetzgebung und Rechtsprechung 414 (2007); C. Morgenstern, ‘Neues zur Führungsaufsicht’, 18 Neue Kriminalpolitik 153 (2006). They bring the example of a high-risk offender who is sanctioned for a violent or a sex offence with imprisonment of, e.g. two years and placed in preventive detention afterwards. Because of the high risk of re-offending, the convicted person will not get an early release after sixteen months (i.e. two-thirds of twenty-four months), but he will have to serve the full term of two years (sec. 57(1) German Criminal Code). Afterwards, he may have to spend up to ten years in preventive detention (sec. 67d(3) German Criminal Code), followed by a period of up to five years during which he will be subject to supervision (sec. 68c(1) German Criminal Code). The convicted person is thus under the control of the criminal justice authorities for up to seventeen years, although his punishment was only two years; this is considered disproportionate.
  The problem with examples like this is that they obtain their persuasive power only from the severity of offence committed in the past, but do not give any information on the seriousness and likelihood of the future offences. Because the duration of preventive detention has to be suspended as soon as it can be expected that the offender will not commit further offences if released (sec. 67d(2) German Criminal Code), the example tacitly implies that there is still a certain risk of re-offending at the end of the ten-year period of preventive detention that excludes an earlier release, but it is, on the other hand, not sufficient for a prolongation of preventive detention for more than ten years. The individual risk may, however, be sufficient to justify the measure of supervision because supervision is a non-custodial measure and therefore less intrusive than preventive detention. With respect to its aim to protect the public against further offences, there are no surrogates in German law capable to achieve the same aim by a smaller burden on the convicted person’s liberty rights. As a result, it seems inevitable to accept that, in the given example, supervision probably is not a disproportionate means of control.⁵²x Ruderich, above n. 32, at 90.
  The example once again shows that the potential of the principle of proportionality is limited because it does not provide for the difficulties to predict reliably dangers to the public. Danger and necessity are concepts that can be described and applied in theory better than in practice. What remains is unease about a preponderance in public policy of security considerations over liberty rights. In this context, it seems worthwhile to note that Kinzig and Baur made a highly interesting suggestion that aims at a legislative re-evaluation of the competing interests. The authors recommend a reduction in the maximum duration of supervision from presently five years (sec. 68c(1) German Criminal Code) to three years, which may be extended for two more years if necessary.⁵³x Baur and Kinzig, above n. 38, at 592. As the diverging interests at hand, public protection as well as liberty rights, are provided for by this suggestion in a balanced and responsible way, it cannot be excluded that the proposition might be taken up by German public policy.

• 5 The Human Rights Perspective

  The compromise between public security and individual liberty rights that defines the convicted person’s legal position in Germany today is only indirectly influenced by the European human rights perspective. The German Federal Constitutional Court ruled in its judgement of 4 May 2011 that the European Convention on Human Rights does not rank higher or on the same level as German constitutional law but that it ranks below the constitution and has only the legal force of ordinary federal law.⁵⁴x Federal Constitutional Court of Germany, 4 May 2011, 2 BvR 2365/09, n. 83. A violation of European human rights can therefore not be contested by an appeal to the German Constitutional Court. The German constitution, i.e. the Basic Law, is nevertheless considered as open and friendly to international law. ‘Friendly’ means that the constitution has to be interpreted in light of the European Convention and the European Court’s judgements as long as this does not lead to a weakening of the standard of protection of the basic rights according to the constitution. The German Constitutional Court identified as one of the examples for such a weakening the so-called multipolar settings involving more than one affected party, where the rights involved needed to be carefully balanced against each other and where more protection given to one party would inevitably imply less protection for the other.⁵⁵x Ibid., n. 93; B. Peters, ‘Germany’s Dialogue with Strasbourg: Extrapolating the Bundesverfassungsgericht’s Relationship with the European Court of Human Rights in the Preventive Detention Decision’, 13 German Law Journal 182 (2012). According to the Constitutional Court, the principle of proportionality is the preferential method to incorporate the European Court’s deliberations into national law.
  In the 1980s, the German Constitutional Court ruled that supervision imposed by law as an automatic consequence of release after two years of imprisonment was consistent with the principle of proportionality.⁵⁶x Federal Constitutional Court of Germany, 15 August 1980, 2 BvR 495/80, n. 6. In those days, the European Convention was not paid much attention to by the German judiciary, though, and the duration of supervision was limited to a maximum of five years. The question thus arises as to whether the law which was expanded in 1998 and 2007 with the option of indeterminate supervision is still in accordance with European fundamental rights, even if these only indirectly influence the interpretation and practice of German national law.
  For such a re-examination, it is necessary to decide in the first step whether supervision is a form of deprivation of the convicted person’s liberty or whether it is just a restriction of liberty. From the German perspective, the distinction may seem to be only of minor importance because the specific safeguards for deprivation of liberty in Article 104 Basic Law (namely formal law and judicial order) are undoubtedly met, but Article 5 ECHR constitutes additional safeguards, which must be met if deprivation of liberty is under scrutiny.⁵⁷x R. Esser, Löwe-Rosenberg. Die Strafprozessordnung und das Gerichtsverfassungsgesetz, 26th edn (2012), EMRK Art. 5 n. 24. In the Guzzardi case, the European Court specified some of the criteria relevant to the distinction between deprivation and restriction of liberty, above all type, duration, effects and manner of implementation of the particular measure.⁵⁸x ECtHR, Guzzardi v. Italy, Applic. No. 7367/76, n. 92. Of these, duration is a criterion that might give cause to classify supervision as deprivation of liberty, because the preventive measure may in exceptional cases last for life, while all other criteria suggest that the convicted person is not deprived of but only restricted in his liberty. The duration of a measure alone, however, may hardly be sufficient to exclusively decide on the appropriate classification of the measure, rather it seems necessary to make an overall assessment of the given criteria. Such an overall assessment would focus on the function of the measure and emphasise the fact that supervision serves as a non-custodial means in the transition process between custody and unrestricted liberty. From the functional point of view, supervision as an abstract sanction category should therefore rather be classified as merely a restrictive measure, which would have the consequence that the safeguards of Article 5 ECHR do not apply.
  A different evaluation might come to mind if supervision is seen in connection with particular directions by which the offender is de facto prevented from leaving certain places or areas, e.g. the direction to carry the equipment for electronic monitoring and not leave home.⁵⁹x Cf. Esser, above n. 57, EMRK Art. 5 n. 27. If these cases are judged as deprivation of liberty, they have to comply with the safeguards of the Convention, i.e . the direction must be considered either as a consequence of the earlier conviction (Art. 5(1)a ECHR) or as necessary to prevent the convicted person from committing further offences (Art. 5(1)c ECHR). Both options have their own disadvantages. In the first case, it seems doubtful whether there is a ‘sufficient causal connection between the conviction and the deprivation of liberty at issue’,⁶⁰x ECtHR, Haidn v. Germany, Applic. No. 6587/04, n. 75. and in the second case, it is doubtful whether the potential further offences are ‘sufficiently concrete and specific’.⁶¹x ECtHR, M. v. Germany, Applic. No. 19359/04, n. 102; Haidn v. Germany, Applic. No. 6587/04, n. 90. So far, the problem has not been discussed in Germany.⁶²x Cf. H.-U. Paeffgen, Systematischer Kommentar zur Strafprozessordnung, 4th edn (2012), EMRK Art. 5 n. 33c. Nevertheless, it can be assumed that the connection between the conviction and the said directions probably is sufficient enough to comply with the Convention. In those cases in which supervision is the automatic consequence of release from prison or a custodial preventive measure, the courts’ subsequent decisions are predetermined by law; they are dependent on the earlier conviction and the determination of the concrete form of punishment.
  If supervision is considered a restriction of liberty, another fundamental right comes into view that is guaranteed in German national law as well as in European law: the right not to be punished twice (Art. 103(3) Basic Law, Art. 4 ECHR Protocol No. 7). In most cases, preventive supervision emerges as a criminal sanction at a time when the primary sanction (imprisonment and/or a custodial preventive measure) has been executed and the convicted person expects to regain unrestricted liberty. From the released offender’s point of view, it can be presumed that the period of guidance and control that follows, if he is under supervision, has the appearance of a second punishment. Nevertheless, technically a violation of the basic right not to be punished twice must be rejected, first because supervision is not considered a form of punishment in Germany (but a preventive measure), and second because it is not construed as an independent judgement, but as a legal, i.e. automatic consequence of the conviction. The judicial decisions that follow after release (e.g. on the duration of supervision, the directions and the cooperation of the institutions) modify the execution of supervision, they are subsequent decisions, but they are not a second punishment for the offence.⁶³x Nuremberg Higher Regional Court, 10 November 2014, 2 Ws 509/14, n. 17. Neither does supervision, by the way, violate the right that there must be no punishment without law (Art. 103(2) Basic Law, Art. 7(1) ECHR) because the specific consequences of supervision are laid down in the Criminal Code and are part of the legal framework that establishes the consequences of the conviction, from the duty to bear the costs of the proceedings to the entry of the conviction and other particulars in the federal central criminal register. If the European Court ruled in the Kafkaris case that the term ‘law’ implies qualitative requirements, including those of accessibility and foreseeability,⁶⁴x ECtHR, Kafkaris v. Cyprus, Applic. No. 21906/04, n. 140. these requirements are met by German law beyond doubt.
  Thus in the end, it is the general principle of proportionality, again, which has to be examined, because in German legal thinking any measure by state authorities that is an interference with the general freedom of action granted by German constitutional law (Art. 2(1) Basic Law) must comply with the rule of law principle of which the proportionality of the measure is an integral part. It has already been pointed out that the German Federal Constitutional Court ruled as early as 1980 that preventive supervision is fully consistent with the principle of proportionality because it may be applied only in those cases in which further offences are expected. The flexibility of the principle, which is a consequence of the open and unspecified notions used in its construction (‘necessary’, ‘adequate’), however, leaves little room for a speculation that how the Constitutional Court would evaluate indeterminate supervision today if the case were brought forward. Neither can it be expected that the European human rights perspective would bring a change because the convicted person’s individual rights would still have to be balanced against public security interests. It is highly probable that the Constitutional Court would reaffirm its earlier decision and hold that supervision is still in accordance with the fundamental rights laid down in the German constitution and the European Convention.
  This finding may seem somewhat disillusioning, but it should be noted that it refers to supervision as only an abstract sanction category, defined by law. It does not refer to the execution of supervision in particular cases, especially to the legitimacy of certain subsequent decisions on the directions imposed on the convicted offender after release. The Dresden Higher Regional Court, for instance, recently found that a direction to take up residence at a specific place prescribed by the competent division of the regional court was not covered by the German Criminal Code and was a violation of the released person’s freedom of movement (Art. 11(1) Basic Law, Article 2(1) ECHR Protocol No. 4).⁶⁵x Dresden Higher Regional Court, 5 June 2015, 2 Ws 248/15, n. 16. In Germany, freedom of movement may only be restricted in the way that the convicted person is directed not to leave his place of domicile or a specified area without the permission of the supervising authority (sec. 68b(1) No. 1 German Criminal Code). In the same way, the Dresden Court found that the direction to make an effort to find employment immediately after release is a violation of the released person’s occupational freedom (Art. 12(1) Basic Law, Art. 15 Charter of Fundamental Rights of the European Union).⁶⁶x Ibid., n. 22. He may only be directed to report to an employment agency (sec. 68b(1) No. 9 German Criminal Code), but he is not obliged to accept any employment offered to him by the agency.⁶⁷x Groß, above n. 26, sec. 68b n. 26; Kinzig (2014), above n. 26, sec. 68b n. 14. The Federal Constitutional Court, too, had to decide several cases in which individual rights had to be balanced against public security interests. In the beginning of 2015, it found in an interim decision that the direction to carry the equipment for electronic monitoring (sec. 68b(1) No. 12 German Criminal Code) is a restriction of the offender’s freedom of action (Art. 2(1) Basic Law), which may be demanded of him until the court finally decides on the case.⁶⁸x Federal Constitutional Court of Germany, 22 January 2015, 2 BvR 2095/14, n. 17. On the other hand, some years earlier, it ruled that the direction not to publish right-wing ideas for a period of five years after release (sec. 68b(1) No. 4 German Criminal Code) was a disproportionate interference with the offender’s freedom of expression (Art. 5(1) Basic Law, Art. 10(1) ECHR).⁶⁹x Federal Constitutional Court of Germany, 8 December 2010, 1 BvR 1106/08. Summing up, it must be concluded that the human rights perspective certainly plays a role in German judicature, but that this role is revealed mainly in connection with the execution of supervision, not in connection with its imposition by law or its function as an abstract sanction category.

• 6 Conclusion

  The indeterminate control of ‘dangerous’ offenders in the community is in Germany provided for by the preventive measure of supervision. Supervision is a combination of control and social support that starts off when a convicted person is released from prison or a custodial preventive measure. Regularly the period of supervision lasts for five years, but it can be shortened as well as extended to indeterminate supervision. In addition, supervision can be combined with directions given to the convicted person to influence his conduct of life and make sure that no further offences are committed. As a result of the legal developments since 1998, the constraints for the imposition and execution of supervision are lower in the case of convicted sex offenders than they are in other cases. While supervision is regularly put into operation, if a convicted person is released from prison after having fully served for at least two years, it is only one year in the case of sex offenders (sec. 68f(1) German Criminal Code). While the requirements for indeterminate supervision are generally quite strict (sec. 68c(2) and (3) German Criminal Code), the period of supervision may be extended for life in the case of convicted sex offenders (and a second, very small group of offenders) if the offender has been sentenced to two years of imprisonment and after release either violates a direction or ‘other specific circumstances give reason to believe that there may be a danger to the general public by the commission of further serious offences’ (sec. 68c(3) No. 2 German Criminal Code). This gives the courts a wide discretion. Against this background one might say that in Germany, too, the legal position of convicted sex offenders is quite exceptional.
  While in the vast majority of cases supervision is imposed by law in Germany, i.e. it is the automatic, legal consequence of release from specified custodial sanctions, the execution of supervision with respect to duration and directions is mainly governed by the principle of proportionality. Suitability, necessity and adequacy are the criteria that are of overall importance in this context. All restrictions imposed upon the convicted person are justified if, and to the degree that, they are suitable, necessary and adequate to achieve the aim of public protection, i.e. to prevent the convicted person from committing further offences. The central legal constraint of proportionality is thus dependent on the validity and reliability of the predictions about the offender’s future criminal conduct, which is in its essence an empirically based statement built on the findings of criminological research. Its reliance on the accuracy of empirical prediction methods delimits the protective function of the proportionality principle to guide and control judicial decisions in the field of preventive sanctions. What is expected by the law, clear and unambiguous assessments on the likelihood and the seriousness of future offences, can, however, hardly be provided for in practice, neither by the courts nor by psychiatric, psychological or criminological expert witnesses. The consequence is in practice a somewhat cloudy mixture of non-distinctive considerations that are flawed by a systematic bias in that they overestimate the risk of re-offending. On the long run, this situation can only be changed by two developments: a clearer distinction between the different target groups of supervision and their differential risks on the side of the courts, and on the side of the experts more empirical research aiming at the further improvement of the validity and accuracy of the prediction methods. The conflict between public security and individual liberty rights, which is in the background of the imposition and the execution of supervision, can only be solved adequately by empirically based, value-oriented judgements of the courts, assigning the diverging interests their due weight.

Noten

• * This article is part of a comparative research project involving several European jurisdictions, of which the results are presented in this special issue. For further background of the theme and outline of the issue and individual articles, as well as the comparative analysis, see the contribution by Van der Wolf in this issue.

• 1 Act to combat sexual offences and other dangerous offences of 26 January 1998, Federal Law Gazette I page 160; the act is commented by H. Schöch, ‘Das Gesetz zur Bekämpfung von Sexualdelikten und anderen gefährlichen Straftaten vom 26.1.1998’, 51 Neue Juristische Wochenschrift 1257 (1998).

• 2 J. van Dijk, J. van Kesteren & P. Smit, Criminal Victimisation in International Perspective. Key Findings from 2004 – 2005 ICVS and EU ICS (2007), at 147; K.-H. Reuband, ‘Steigende Repressionsneigung im Zeitalter der “Postmoderne”?’, 15 Neue Kriminalpolitik 100 (2003); K.-H. Reuband, ‘Steigende Punitivität oder stabile Sanktionsorientierungen der Bundesbürger? Das Strafverlangen auf der Deliktebene im Zeitvergleich’, 21 Soziale Probleme 97 (2010).

• 3 E.C. Rautenberg, ‘Wegschließen für immer!?’, 54 Neue Juristische Wochenschrift 2608 (2001).

• 4 J. Kinzig, Die Legalbewährung gefährlicher Rückfalltäter. Zugleich ein Betrag zur Entwicklung des Rechts der Sicherungsverwahrung (2008), at 9.

• 5 Act to reform supervision and to revise the provisions on subsequent preventive detention of 13 April 2007, Federal Law Gazette I page 513, with commentary by U. Schneider, ‘Die Reform der Führungsaufsicht’, 27 Neue Zeitschrift für Strafrecht 441 (2007).

• 6 ECtHR, M. v. Germany, Applic. No. 19359/04.

• 7 ECtHR, Haidn v. Germany, Applic. No. 6587/04.

• 8 Federal Constitutional Court of Germany, 4 May 2011, 2 BvR 2365/09.

• 9 Act to reorder the provisions on preventive detention of 22 December 2010, Federal Law Gazette I page 2300.

• 10 Act to implement the imperative of distance in the provisions on preventive detention of 5 December 2012, Federal Law Gazette I page 2425.

• 11 Council of Europe, Annual Penal Statistics. Space I – Prison Populations. Survey 2013 (2014), at 42.

• 12 J.-M. Jehle, H.-J. Albrecht, S. Hohmann-Fricke & C. Tetal, Legalbewährung nach strafrechtlichen Sanktionen. Eine bundesweite Rückfalluntersuchung 2007 bis 2010 und 2004 bis 2010 (2013), at 36, available at: <www.bmjv.de/DE/Ministerium/Fachthemen/AbtII/IIA7/Rueckfallstatistik_doc.html?nn=1468684> (last visited 12 August 2015). An overview of recidivism studies in various European countries is given on the Internet: <http://wp.unil.ch/space/publications/recidivism-studies/27/> (last visited 19 April 2016).

• 13 Statistisches Bundesamt, Rechtspflege. Strafverfolgung (2013), table 5.5.

• 14 B.-D. Meier, Strafrechtliche Sanktionen, 4th edn (2015), at 296-301.

• 15 See <www.dbh-online.de/fa/> (last visited 31 July 2015).

• 16 M.P. Rohrbach, ‘Die Verschärfung der Führungsaufsicht’, 97 Monatsschrift für Kriminologie und Strafrechtsreform 248 (2014); see also C. Morgenstern, A. Hecht, ‘Rechtstatsachen zur Führungsaufsicht im kriminalpolitischen Kontext’, 58 Bewährungshilfe 186 (2011).

• 17 T. Vormbaum, Einführung in die moderne Strafrechtsgeschichte, 2nd edn (2011), at 137.

• 18 Reichsjustizamt, Vorentwurf zu einem Deutschen Strafgesetzbuch. Begründung. Allgemeiner Teil (1909), at 147.

• 19 Act against dangerous habitual offenders and on measures of security and rehabilitation of 24 November 1933, Imperial Law Gazette I page 995.

• 20 Federal Court of Justice, 18 March 1952, GSSt 2/51.

• 21 Meier, above n. 14, at 192; see also F. Streng, Strafrechtliche Sanktionen, 3rd edn (2012), at 255.

• 22 Federal Court of Justice, 4 August 1965, 2 StR 282/65.

• 23 W. Frisch, ‘Die Maßregeln der Besserung und Sicherung im strafrechtlichen Rechtsfolgensystem’, 102 Zeitschrift für die gesamte Strafrechtswissenschaft 369, at 382 (1990); G. Kaiser, Befinden sich die kriminalrechtlichen Maßregeln in der Krise? (1990), at 12; see also Federal Constitutional Court of Germany, 4 May 2011, 2 BvR 2365/09.

• 24 A. Dessecker, Gefährlichkeit und Verhältnismäßigkeit. Eine Untersuchung zum Maßregelrecht (2004), at 132.

• 25 Ibid., at 187.

• 26 Marburg Regional Court, 13 January 2014, 11 StVK 9/14; K.-H. Groß, Münchener Kommentar zum Strafgesetzbuch, 2nd edn (2012), sec. 68c n. 7; J. Kinzig, Schönke/Schröder. Strafgesetzbuch. Kommentar, 29th edn (2014), sec. 68c n. 2.

• 27 T. Hörnle, Tatproportionale Strafzumessung (1999); T. Hörnle, ‘Das antiquierte Schuldverständnis der traditionellen Strafzumessungsrechtsprechung und –lehre’, 54 Juristenzeitung 1080 (1999); but see also Streng, above n. 21, at 314-17.

• 28 Streng, above n. 21, at 166-67; W. Frisch, ‘Defizite empirischen Wissens und ihre Bewältigung im Strafrecht’, in R. Bloy, M. Böse, T. Hillenkamp, C. Momsen & P. Rackow (eds.), Gerechte Strafe und legitimes Strafrecht. Festschrift für Manfred Maiwald (2010), 239 at 252; W. Frisch, Prognoseentscheidungen im Strafrecht. Zur normativen Relevanz empirischen Wissens und zur Entscheidung bei Nichtwissen (1983), at 53-55, 161-70.

• 29 J. Kaspar, Verhältnismäßigkeit und Grundrechtsschutz im Präventionsstrafrecht (2014), at 101; Dessecker, above n. 24, at 343.

• 30 Federal Constitutional Court of Germany, 25 February 1975, 1 BvF 1/74; 29 May 1993, 2 BvF 2/90; K. Stern, ‘Die Schutzpflichtenfunktion der Grundrechte: Eine juristische Entdeckung’, 63 Die öffentliche Verwaltung 241 (2010); V. Epping, Grundrechte, 6th edn (2015), at 59.

• 31 A. Dessecker, ‘Von der Polizeiaufsicht zur Führungsaufsicht und zurück?’, in E. Hilgendorf and R. Rengier (eds.), Festschrift für Wolfgang Heinz (2012) 631, at 634.

• 32 D. Ruderich, Führungsaufsicht (2014), at 245; S. Thomaßen, ‘Konzeption zum Umgang mit rückfallgefährdeten Sexualstraftätern (KURS NRW)’, 6 Forensische Psychiatrie, Psychologie, Kriminologie 25 (2012); M.P. Rohrbach, Die Entwicklung der Führungsaufsicht unter besonderer Berücksichtigung der Praxis in Mecklenburg-Vorpommern (2014), at 152; Rohrbach, above n. 16, at 257.

• 33 F. Schoch, ‘Doppelfunktionale Maßnahmen der Polizei’, 35 Juristische Ausbildung 1115 (2013).

• 34 T. Siegel, ‘Hooligans im Verwaltungsrecht. Stadionverbote und andere polizeirechtliche Maßnahmen zur Eindämmung von Gewalt in Fußballstadien’, 66 Neue Juristische Wochenschrift 1035 (2013).

• 35 ECHR, Ostendorf v. Germany, Applic. No. 15598/08.

• 36 H. Schneider, Strafgesetzbuch. Leipziger Kommentar, 12th edn (2008), sec. 68d n. 6.

• 37 For determinate supervision, see Federal Constitutional Court of Germany, 15 August 1980, 2 BvR 495/80.

• 38 A. Baur and J. Kinzig (eds.), Die reformierte Führungsaufsicht (2015).

• 39 Dresden Higher Regional Court, 12 December 2008, 2 Ws 380/08; H. Ostendorf, Nomos Kommentar Strafgesetzbuch, 4th edn (2013), sec. 68c n. 5; but see also Berlin Higher Regional Court, 20 June 2011, 2 Ws 159/11.

• 40 Baur and Kinzig, above n. 38, at 533.

• 41 Ibid., at 538 and 543.

• 42 Jehle et al., above n. 12, at 83.

• 43 J.L. Müller, ‘Legalbewährung nach Gutachten zur nachträglichen Sicherungsverwahrung’, 94 Monatsschrift für Kriminologie und Strafrechtsreform 256 (2011).

• 44 M. Alex, ‘Kriminalprognose und Legalbewährung – Wie zuverlässig lässt sich Rückfallgefahr vorhersagen?’, in K. Höffler (ed.), Brauchen wir eine Reform der freiheitsentziehenden Sanktionen? (2015) 21, at 26.

• 45 Kinzig (2008), above n. 4, at 213 and 217.

• 46 J. Kühl and K.F. Schumann, ‘Prognosen im Strafrecht – Probleme der Methodologie und Legitimation’, 7 Recht & Psychiatrie 126 (1989).

• 47 Meier, above n. 14, at 372.

• 48 Alex, above n. 44, at 33.

• 49 Act on the therapeutic placement of mentally disturbed violent offenders of 22 December 2010, Federal Law Gazette I page 2300.

• 50 Federal Constitutional Court of Germany, 11 July 2013, 2 BvR 2302/11.

• 51 Ostendorf, above n. 39, pre secs. 68-68g n. 15; see also H. Pollähne, ‘Effektivere Sicherheit der Bevölkerung und schärfere Kontrolle der Lebensführung’, 90 Kritische Vierteljahresschrift für Gesetzgebung und Rechtsprechung 414 (2007); C. Morgenstern, ‘Neues zur Führungsaufsicht’, 18 Neue Kriminalpolitik 153 (2006).

• 52 Ruderich, above n. 32, at 90.

• 53 Baur and Kinzig, above n. 38, at 592.

• 54 Federal Constitutional Court of Germany, 4 May 2011, 2 BvR 2365/09, n. 83.

• 55 Ibid., n. 93; B. Peters, ‘Germany’s Dialogue with Strasbourg: Extrapolating the Bundesverfassungsgericht’s Relationship with the European Court of Human Rights in the Preventive Detention Decision’, 13 German Law Journal 182 (2012).

• 56 Federal Constitutional Court of Germany, 15 August 1980, 2 BvR 495/80, n. 6.

• 57 R. Esser, Löwe-Rosenberg. Die Strafprozessordnung und das Gerichtsverfassungsgesetz, 26th edn (2012), EMRK Art. 5 n. 24.

• 58 ECtHR, Guzzardi v. Italy, Applic. No. 7367/76, n. 92.

• 59 Cf. Esser, above n. 57, EMRK Art. 5 n. 27.

• 60 ECtHR, Haidn v. Germany, Applic. No. 6587/04, n. 75.

• 61 ECtHR, M. v. Germany, Applic. No. 19359/04, n. 102; Haidn v. Germany, Applic. No. 6587/04, n. 90.

• 62 Cf. H.-U. Paeffgen, Systematischer Kommentar zur Strafprozessordnung, 4th edn (2012), EMRK Art. 5 n. 33c.

• 63 Nuremberg Higher Regional Court, 10 November 2014, 2 Ws 509/14, n. 17.

• 64 ECtHR, Kafkaris v. Cyprus, Applic. No. 21906/04, n. 140.

• 65 Dresden Higher Regional Court, 5 June 2015, 2 Ws 248/15, n. 16.

• 66 Ibid., n. 22.

• 67 Groß, above n. 26, sec. 68b n. 26; Kinzig (2014), above n. 26, sec. 68b n. 14.

• 68 Federal Constitutional Court of Germany, 22 January 2015, 2 BvR 2095/14, n. 17.

• 69 Federal Constitutional Court of Germany, 8 December 2010, 1 BvR 1106/08.